A Case of Confounding Back Pain

Background

Acute back pain remains one of the most common prompts for medical evaluation [1], with an annual prevalence estimated at 15% to 20% in the United States and a lifetime prevalence of more than 60% in primary care settings [2]. In the absence of red flags, such as fever, injury, bowel or bladder dysfunction, or neurologic deficits, most cases can be treated conservatively without the need for imaging studies, and spontaneous resolution of symptoms can be expected [2–4]. Indeed, the Choosing Wisely campaign, supported by the American Board of Internal Medicine and American Academy of Family Physicians, recommends no imaging within the first 6 weeks of symptoms in patients with nonspecific back pain without red flag symptoms [5,6]. However, maintaining a broad differential diagnosis and encouraging close follow-up remains vital. One of the more concerning causes of back pain with red flag signs is discitis, inflammation of the intervertebral space, which is often caused by infection and accompanied by osteomyelitis of the surrounding bone [7,8]. The most common causes of infectious discitis/osteomyelitis are hematogenous spread of organisms (rather than contiguous spread) and direct inoculation from trauma or invasive procedures [7]. We present the case of a patient with back pain, which was most likely caused by hematogenous spread of bacteria. Magnetic resonance imaging (MRI) of the spine has high diagnostic value in cases of discitis/osteomyelitis, and needle biopsy with bacterial culture guides antibiotic treatment.

Case Report

HISTORY:

A 70-year-old man was admitted to the hospital for 5 days with a diagnosis of sepsis in the setting of a urinary tract infection (UTI) with multidrug-resistant (MDR) and extended-spectrum β-lactamase (ESBL)-producing Escherichia coli, along with bacteremia. He was previously healthy and had a history of hypertension and benign prostatic hyperplasia (BPH). Repeated blood cultures obtained in the hospital 24 h after initiation of antibiotics were negative. After being deemed clinically stable, he was discharged home and received intravenous (i.v.) ertapenem 1 g daily for 2 weeks, with the assistance of home health care. Furthermore, during this hospital admission, he had development of diarrhea and was found to have Clostridioides difficile colitis, which was treated with oral vancomycin 125 mg every 6 h for 10 days. During a follow-up visit at his outpatient family medicine clinic 6 days after hospital discharge, he reported ongoing back pain since discharge. At that time, he had no neurologic deficits. Therefore, the pain was suspected to be musculoskeletal, potentially due to immobility while hospitalized. He was treated conservatively with a recommendation for over-the-counter acetaminophen, heat therapy, and lidocaine patches.

At his second follow-up visit 2 weeks after discharge, the patient again reported ongoing back pain that was not responding to conservative management. The patient was instructed to begin physical therapy, and a radiograph of the thoracic spine was obtained. This radiograph showed only multilevel discogenic degenerative changes along with diffuse idiopathic skeletal hyperostosis in the mid and lower thoracic spine. The patient had a third outpatient appointment 10 days later because his pain still had not improved with conservative management, including physical therapy. Follow-up thoracic spine MRI with and without i.v. contrast, performed 2 weeks after radiography, showed T7-T8 osteomyelitis and discitis with significant paraspinal soft tissue involvement, along with right-sided T6-T7 paracentral disc extrusion with mild right-sided hemicord compression and a probable T5 hemangioma (Figure 1A). Given these findings, the patient was urgently referred to the Emergency Department for admission.

Upon readmission, the patient appeared nontoxic and was hemodynamically stable. He continued to endorse thoracic back pain, and his examination was significant for pain on palpation of the T7 and T8 regions. Initial laboratory studies were significant for leukocytosis (leukocytes, 11×10⁹/L; reference range, 3.5–10.5×10⁹/L), hyponatremia (sodium, 134 mEq/L; reference range, 136–145 mEq/L), and a lactate level of 2.3 mmol/L (reference range, <2.2 mmol/L), which resolved to 1.1 mmol/L after fluid resuscitation. His initial C-reactive protein (CRP) level was high (75 mg/L; reference range, <10 mg/L) but normalized to less than 3 mg/L toward the end of the hospital course. His initial glomerular filtration rate was greater than 90 mL/min/1.73 m², and no subsequent decreases in kidney function occurred at any point during his treatment. Repeated blood cultures during this admission were negative. Surgical intervention was not recommended, owing to the absence of neurologic deficits. He underwent computed tomography (CT)-guided biopsy of the lesion at T7–T8 and was started on empiric i.v. ertapenem 1 g daily and i.v. vancomycin 1500 mg twice daily. Culture of the biopsy was significant for MDR and ESBL-producing E. coli, which was found to be sensitive to ertapenem, similar to the urine and blood culture results from the prior admission. Therefore, the vancomycin was discontinued. The patient underwent peripherally inserted central catheter placement for a planned prolonged period of i.v. ertapenem administration, 1 g daily for 8 weeks.

After 6 days in the hospital, the patient was discharged home and noted an initial improvement in his back pain with the daily ertapenem. However, he returned to the outpatient clinic 4 weeks later with worsening back pain. MRI of the thoracic spine was performed again, which showed progressive discitis/osteomyelitis along with phlegmonous changes and a small, left-sided paraspinal abscess (Figure 1B). He was readmitted to the hospital and underwent CT-guided aspiration of the T7–T8 discitis/osteomyelitis lesion. The ertapenem was discontinued, and the patient was started on empiric i.v. vancomycin 1500 mg twice daily and piperacillin-tazobactam 3.375 g every 6 h for concerns of ertapenem failure. Again, no neurosurgical intervention was recommended owing to a lack of neurologic deficits. The patient noted improvement after aspiration of the abscess.

Bacterial culture of the CT-guided aspiration again indicated the same MDR E. coli, with resistance to ampicillin, cefazolin, cefepime, ceftriaxone, and ceftazidime (Figure 2). The cultured specimen also was not susceptible to tetracycline, levofloxacin, or gentamicin and showed only intermediate susceptibility to piperacillin-tazobactam (minimum inhibitory concentration, 16 µg/mL). Because the E. coli showed minimal growth with trimethoprim as well as ertapenem, the patient was switched back to i.v. ertapenem 1 g daily from i.v. vancomycin/piperacillin-tazobactam. The clinical improvement was attributed to better source control from the aspiration rather than from the empiric use of vancomycin/piperacillin-tazobactam. After 5 days of hospital admission, he was discharged again with another full 8-week course of i.v. ertapenem 1 g daily, followed by a 4-week course of oral trimethoprim-sulfamethoxazole (160–800 mg), 1 tablet twice daily, per infectious disease specialist recommendations.

POSTHOSPITAL COURSE:

The patient continued to be followed up in the outpatient primary care and infectious disease clinics. His back pain gradually improved, and he noted substantial improvement at 12 weeks after discharge. Subsequent MRI of the thoracic spine (Figure 1C) showed overall improvement of the T7-T8 discitis/osteomyelitis, along with a decrease in surrounding soft tissue inflammation. He was enrolled in physical therapy to help support his recovery. Eight months after initiating physical therapy, MRI of the thoracic spine noted interval resolution of T7 and T8 osteomyelitis (Figure 3). The patient reported steady improvement in his back pain, and he was agreeable to continue with physical therapy.

Discussion

Vertebral osteomyelitis (VO) is rare, occurring in 1 in 250 000 to 450 000 persons per year, and with an incidence of only 0.02% of admissions over 10 years [9]. Furthermore, in 1 study, Staphylococcus aureus was the most common cause of VO, accounting for 51% (38/74) of the reported cases [9]. The incidence of VO in the United States is increasing, most likely due to the increasing age of the population and improved detection methods [10]. VO has been associated with several risk factors, including diabetes, immunosuppression, advanced age, long-term use of corticosteroids, and spinal surgery [11–13]. It is important to remember that patients with these risk factors are at risk for hematogenous spread to the bones when they have bacteremia [14].

In the case of our patient, who had no clear immunosuppression, the only identifiable risk factor for bacteremia and sepsis that eventually led to VO was a UTI in the setting of BPH. BPH causing bladder outlet obstruction has been reported to increase the incidence of lower UTI in men [15,16]. Therefore, it is important to treat obstructive BPH because this can prevent urine retention and reduce the risk of UTI, which ultimately would have prevented the original bacteremia in the case of this patient.

For patients with considerable suspicion for VO, especially those with acute back pain and red flag signs (eg, fever, spinal tenderness, and neurologic dysfunction) [17], any normal radiography findings should prompt further evaluation, such as with MRI. Although back pain has a broad differential diagnosis, fever indicates possible infection, which occurs in approximately 35% to 60% of cases [12]. However, because fever and back pain can indicate other processes, such as infection of the retroperitoneum [12] or pyelonephritis, testing for VO should include MRI after radiography. Infectious Diseases Society of America (IDSA) guidelines [18] strongly recommend MRI of the spine as the diagnostic imaging of choice for suspected VO because of its sensitivity of 97%, specificity of 93%, and accuracy of 94% in diagnosing VO. For patients with back pain in a setting of bacteremia, clinicians’ index of suspicion for VO should be higher, and the threshold for the use of MRI for diagnosis should be low. For patients with suspected VO who have multiple focal points of tenderness on spine examination, full spine MRI can be warranted to assess for multiple foci of infection. Physical examination in our patient indicated point tenderness at the thoracic spine, which prompted thoracic spine MRI with and without contrast.

Laboratory studies for the evaluation of VO should include a complete blood cell count with differential, CRP measurement, and blood culture [12]. If laboratory findings are abnormal, the patient should be considered for bone biopsy to assist in targeted antibiotic medical management [10,12]. IDSA guidelines [18] stress the importance of assessing inflammatory markers, such as CRP and erythrocyte sedimentation rate. Although these markers are nonspecific, if increased in patients with back pain, they have 94% to 100% sensitivity for VO. Therefore, these tests are strongly recommended for monitoring of prognosis, especially after 4 weeks of antimicrobial therapy. CRP is reportedly more useful than erythrocyte sedimentation rate because it becomes increased much earlier in a disease course [5]. As seen in the current case, up to 40% of patients with VO have a normal leukocyte count [18]. Because the patient had a nontoxic appearance and was hemodynamically stable, a detailed physical examination and necessary laboratory and imaging studies were critical for an accurate diagnosis.

IDSA guidelines [18] report S. aureus as the main microbiologic cause of VO and state that blood cultures can be inadequate in the diagnosis of VO. Therefore, if a patient is clinically suspected of having VO, but blood cultures do not identify S. aureus, Staphylococcus lugdunensis, or Brucella species, IDSA recommends image-guided biopsy as the next step [18]. The current case exemplified this situation. Our patient had negative blood cultures after 24 h of antimicrobial treatment, but hematogenous spread of E. coli to the spine still occurred. The persistence of back pain despite initial management, in the setting of a previous diagnosis of systemic infection, was an important clue to identifying VO as a potential diagnosis in this patient. With the atypical disease presentation and the finding of bacteremia not due to S. aureus, S. lugdunensis, or Brucella, this patient was a good candidate for image-guided aspiration biopsy. This was key in guiding his diagnosis and therapy.

The IDSA recommendations for treatment are based on ensuring that therapy is directed toward the cause [18]. Therapy can even be held until a microbiologic diagnosis is established, as long as the patient is stable. It is important, however, that the total duration of antibiotics be 6 weeks, which can be extended if there is concern for persisting infection [18]. The exception is that Brucella species should be treated for 3 months. In the case of our patient, the biopsy results showed that he had MDR and ESBL-producing E. coli, and his antibiotics were adjusted appropriately on the basis of his culture results. He was also treated for more than 6 weeks per recommendations from our infectious disease specialists.

It is important to remember that standard radiography findings can be negative early in osteomyelitis, which further solidifies the importance of post-hospital follow-up for all patients within 1 or 2 weeks after hospital discharge. This is especially true for those with bacteremia while in the hospital. If conservative treatment including physical therapy does not improve pain, other causes such as infection should be considered.

Conclusions

Acute back pain has a broad differential diagnosis, including VO. Systemic infections pose a substantial risk for VO, especially in older and immunocompromised patients. The infection can seed hematogenously, with distant sites (eg, urinary tract and skin) acting as focal sources of infection, and can spread to the vertebrae, causing osteomyelitis. Patients with acute back pain require a detailed assessment in terms of a complete history and physical examination with a high index of suspicion for VO in the presence of red flag signs. Appropriate imaging and laboratory investigations are also necessary.

A patient who returns to the clinic with back pain that has not improved should always be reevaluated for the possibility of a diagnosis different from the initially perceived cause. A patient with persistent back pain after known recent bacteremia can benefit from early imaging and screening CRP measurement. Even if the initial radiography findings are negative, clinical evaluation can indicate the importance of MRI to assess for the possibility of VO and/or discitis. In the case of our patient, MRI performed within a few weeks of conservative management was important for preventing further progression of the initially confounding infection.